Are digital-using UK firms more productive?


Are digital-using UK firms more productive?

By Diane Coyle, Kieran Lind, David Nguyen, Manuel Tong

A large and growing literature addresses two trends in productivity: one is a slowdown in productivity growth in many developed economies from the mid-2000s (Lafond et al 2021); and a second is evidence of a growing productivity gap between frontier firms and the rest (Andrews, Criscuolo & Gal 2019, Haldane 2017). One of the reasons these phenomena call for investigation is that they have emerged during a period when there has been continuing digitalisation of production, reflected for example in rapid increases in data usage and cloud services, and a proliferation of digital platform business models. This might have been expected to increase firm-level and aggregate productivity. 

At least two types of resolution to the puzzle have been suggested, some with conflicting implications. One is that there are fewer new ideas or important innovations, as compared with previous periods of high productivity growth (Bloom et al 2020, Gordon 2017). This line of argument points to productivity growth remaining low despite continuing adoption of new technologies; it considers these to be simply less productive than previous ones.

An alternative is that there are intangible aspects of digital adoption, such as the time needed by most firms to achieve productivity gains, for example because of necessary organisational changes (Tambe et al. 2020). In this case, a small number of firms might record productivity improvements, but the gains would take time to spread to most firms in the economy. Therefore, one possible explanation for the productivity slowdown in advanced economies, coinciding with widespread digital adoption, is that firms need time to change organisational structures or processes to use the new technologies effectively. This organisational delay, combined perhaps with the need to make additional investments or acquire skills and know-how, has been labelled the ‘productivity J-curve’. To explore this, we need to look at the adoption of digital technologies by different individual firms and see whether it is linked to firm-level productivity.

Using a unique UK firm-level data set, in a new ESCoE Discussion Paper published today, we explore the links between a large set of digital inputs and investments and productivity. We assembled the most comprehensive database to date on UK firms’ purchases of inputs and investments, and specifically a wide range of digital measures, such as internet access, orders via website, ICT specialists, cloud computing, among others. We have linked for the first time three sets of data on firms’ activities (the Annual Business Survey, Annual Purchases Survey, and E-Commerce Survey), in total comprising around 11,000 firms. We can therefore include a wide range of inputs to estimate production functions that are able to account for the contribution of the adoption of different digital technologies to firm-level productivity.

Our findings based on this unique UK dataset confirm that it is the largest firms (by number of employees) which are on average the most productive. We found that large firms are more digital-intensive than small ones and that digital adopters do have higher productivity than non-adopters, but the nature of the digital variables matters. There is a consistent message that not all digitalisation variables in our data from E-Com can be interpreted in the same way; several of them involving external purchases correlate with lower productivity, possibly because they indicate lower in-house digital capabilities. Having key digital capabilities available in-house, as opposed to purchasing in digital services from external suppliers, seems to be an important indicator. For example, the employment of in-house ICT specialists is the most significant “digital” variable, positively associated to firm productivity, while the use of CRM software, cloud computing and 3D printing also play a role. The use of multiple in-house digital technologies is also strongly positively linked to TFP.

However, other digital variables available in our data set are negatively related to productivity. These are simple zero-one indicators of whether or not the firm makes external purchases of various services such as software support, web solutions and data protection. These may signal an important difference between having in-house capabilities and needing to purchase services from external suppliers.

Our findings therefore support the interpretation that organisational factors internal to the firm, as well as specific employee skills, are key to deriving a productivity advantage from adopting digital technologies, and help explain the gap between leader and laggard firms. This finding that firms’ capabilities matter for the impact of digital adoption on productivity takes advantage of the wide range of digital variables we were able to use, and points to the need for future research on the role of digital technology in driving productivity to take account of organisational capabilities.

Read the full ESCoE Discussion Paper here.

Andrews, D., Criscuolo, C. & Gal, P.N., (2019). ‘The best versus the rest: divergence across firms during the global productivity slowdown’. CEP Discussion Papers, dp1645. Centre for Economic Performance, LSE. Available at:

Bloom, N., Jones, C.I., Van Reenen, J. & Webb, M., (2020). ‘Are ideas getting harder to find?’. American Economic Review110(4), pp.1104-44. Available at:

Gordon, R., (2017). The Rise and Fall of American Growth (pp. 769-770). Princeton University Press.

Haldane, A., (2017). Productivity puzzles. Speech given at the London School of Economics, 20 March. Available at:

Lafond, F., Goldin, I., Koutroumpis, P., & Winkler, J., (2021). ‘Why is productivity slowing down?’, INET Oxford Working Papers 2021-12, Institute for New Economic Thinking at the Oxford Martin School, University of Oxford.

Tambe, P., Hitt, L., Rock, D. & Brynjolfsson, E., (2020). ‘Digital capital and superstar firms,’ NBER Working Paper 28285. National Bureau of Economic Research. Available at:

Diane Coyle is Bennett Professor of Public Policy at the University of Cambridge
Kieran Lind is an Economist at the Office for National Statistics (ONS)
David Nguyen is an Economist at the Organisation for Economic Co-operation and Development (OECD)
Manuel Tong is an Economist at the National Institute of Economic and Social Research (NIESR)

ESCoE blogs are published to further debate. Any views expressed are solely those of the author(s) and so cannot be taken to represent those of the ESCoE, its partner institutions or the Office for National Statistics.

About the authors

Kieran Lind

Manuel Tong Koecklin

Research Projects

Related publications